Jump to content

Flashing cursor after install


Recommended Posts

Here's the situation. We are trying to install POSReady 2009 (for all intents and purposes, this is XP) on a RAID1 on an Intel desktop controller. Setup is normal (either provide iastor driver via click or putting in drivers folder) it detects the HDD, we can format it and install the OS. Checking the OS after setup (in the PE it goes back to command prompt) we can change to C: and see that there are files there.

After a restart, it will not boot off the RAID volume, all we get is a flashing cursor. I have also tried using bootsect on the volume after setup. Boot.ini is correct.

You can probably ignore the fact that this is POSReady, as I am just looking for other things to try or other possible reasons why this would happen on XP.

The RAID driver is in the "image" already and we are not getting a 0x7b... just a flashing cursor.

Booting this into our WDS (WinPE 3.0) we can see the volume has an OS on it, however running Diskpart it shows that the volume is active, but it also shows "Boot Disk: NO"

Any ideas?

PS: no problem on this hardware using IDE mode instead of RAID.

Link to comment
Share on other sites


Here's the situation. We are trying to install POSReady 2009 (for all intents and purposes, this is XP) on a RAID1 on an Intel desktop controller. Setup is normal (either provide iastor driver via click or putting in drivers folder) it detects the HDD, we can format it and install the OS. Checking the OS after setup (in the PE it goes back to command prompt) we can change to C: and see that there are files there.

After a restart, it will not boot off the RAID volume, all we get is a flashing cursor. I have also tried using bootsect on the volume after setup. Boot.ini is correct.

You can probably ignore the fact that this is POSReady, as I am just looking for other things to try or other possible reasons why this would happen on XP.

The RAID driver is in the "image" already and we are not getting a 0x7b... just a flashing cursor.

Booting this into our WDS (WinPE 3.0) we can see the volume has an OS on it, however running Diskpart it shows that the volume is active, but it also shows "Boot Disk: NO"

Any ideas?

PS: no problem on this hardware using IDE mode instead of RAID.

The flashing boot cursor is traditionally linked to:

  1. non-balance between CHS and LBA in the partition table
  2. "wrong" geometry in the bootsector BPB

I doubt the first can be the cause, so I would go for the second.

Try nopping the checks for geometry in the bootsector code.

See here:

http://www.911cd.net/forums//index.php?showtopic=21702&st=129

http://reboot.pro/8528/page__st__21

Alternatively, can you boot off a grub4dos (on CD/USB/floppy whatever)?

jaclaz

Link to comment
Share on other sites

My options for boot are WinPE v1.5, 3.0 (x86 and x64) or DOS 7.1 (which has NTFSDOS). However I thought about this for a moment, while I can boot to DOS I won't be able to load a RAID driver to see the volume...

I also have the ability to boot to our test server which has a PXELINUX rom if gives any other options.

UPDATE: ran a program called MBRFix from the WinPE 2.0 (x86) which fixed it. However the same program (and commands) did not work from the WinPE 1.5 that XPe installs from. It seems to working now. :rolleyes:

Link to comment
Share on other sites

My options for boot are WinPE v1.5, 3.0 (x86 and x64) or DOS 7.1 (which has NTFSDOS). However I thought about this for a moment, while I can boot to DOS I won't be able to load a RAID driver to see the volume...

I also have the ability to boot to our test server which has a PXELINUX rom if gives any other options.

UPDATE: ran a program called MBRFix from the WinPE 2.0 (x86) which fixed it. However the same program (and commands) did not work from the WinPE 1.5 that XPe installs from. It seems to working now. :rolleyes:

Then you had an empty MBR (no code) provided that my crystal ball is accurate in seeing HOW you did run MBRFIX.

BTW you don't need a RAID driver to acces a single sector AFAIK.

jaclaz

Edited by jaclaz
Link to comment
Share on other sites

A blank MBR makes sense. I do understand the "simple" reasons why you get a flashing cursor. As I explained it to my compatriot, the system sees the HDD (or the volume) but doesn't know what to do with it. Otherwise it would go to the "insert a disk" type message.

I'm just curious as to why installing it on IDE mode works fine, but installing it on RAID mode with iaStor.inf wouldn't write the MBR data. Perhaps it is a bug with this version of Windows?

Link to comment
Share on other sites

Here is a strange situation that I've found, and so far have only found it in WinPE v2.1 (x86) and WinPE v3.0 (x64). Other versions may be affected as well.

First a little story of what I am dealing with here.

Or to sum it up. It appears that Windows XP Embedded POSReady2009, while it does support RAID drivers, it doesn't write the MBR if you install it with those drivers. It does write the MBR if you install using a default (IDE mode) driver, which is a problem entirely on its own. The end result is I have captured a POSReady2009 (RAID) image to my WDS. Since it is known that Imagex does not capture MBR code when creating an image, it is entirely possible deploying this image on a brand new (blank) HDD will result in the client not booting.

To test this theory out, I deployed the image onto a notebook, which is currently set for AHCI. i was hoping for a 0x7b error, which I was able to get. HOWEVER, this HDD previously had an OS on it, so its MBR would have data in it already. I was about to test out erasing the MBR so that I can test the RAID image on it, when I found this oddity.

Even though there is only 1 HDD in this system, WinPE (and Diskpart) is reporting that there are two HDDs in the system. Is this normal?

My Gimagex HTA is showing them as such:

C:\system     (Hard Disk)   NTFS   Size:   2GB / 466GB
E: (Usb-Drive) Size: 2GB / 466GB

C: is accessable and I can see the files, but E: is not accessable. Upon attempting to access it, I receive:

There is no disk in the drive. Please insert a disk into drive \Device\Harddisk1\DR1.

Diskpart is even fooled to believe there are two physical disks in the system as well. A LIST DISK returns:

Disk 0        Online        465 GB    0 B
Disk 1 No Media 0 B 0B

When I do a LIST VOL, I can see that E: is of type "Removable" which is certainly correct, as the E: is a card reader! However, once having this RAID enabled image on it, portions of WinPE seem to want to attribute the second portion of the "broken mirror" to the card reader?

Link to comment
Share on other sites

Tripredacus, IMHO you are failing to test or report correctly. (as you did on the linked thread).

As soon as a DIsk (I mean completely 00ed disk) is opened in Disk Management, the Disk Management should ask you to "Initialize" the disk.

When you click YES, this normally results in three DIFFERENT sets of writes to the disk first sector:

  1. the MBR CODE
  2. the Disk Signature
  3. the "magic bytes" signature 55AA

Then when you create a partition on the disk, (let's for the sake of simplicity that you create a Primary partition) a single entry is written to the disk first sector:

  • a partition table entry (or if you prefer MBR data)

Then when you set the partition as "Active" or "Bootable", a single byte is written to the first byte of the partition table entry:

  • 80 (still part of the MBR data)

In other words, if before trying doing other things/different tests, you would simply 00 out frst sector of the disk and repeat the same exact steps you did here:

possibly better describing them in more detail AND making a copy of said first sector after each subsequent "stage" of that procedure, we may have something to work on.

Particularly if you could better describe the actual install procedure you are doing:

Setup is normal (either provide iastor driver via click or putting in drivers folder) it detects the HDD, we can format it and install the OS.

AFAIK there is NOTHING like a "normal" Setup, what you are probably doing is to boot a PE of some kind (possibly a 2.x or a 3.x) and from it install the actual PosReady from it's .iso. (but this are mainly my guesses :ph34r: )

jaclaz

Link to comment
Share on other sites

Ah the reason why I made a separate post was mainly to inquire why putting this image (from a .wim) onto a notebook with 1 HDD installed would end up reporting that the Card Reader was the second drive in the mirror.

As far as the "Setup" it involved making the "ISO" (its not an ISO truly, we have the pressed disc from MS) boot from a USB key. This let us then put the iaStore.inf (and related files) into the appropriate folders as directed by our instructions. This only bypasses the "F6" method (which is just a button similar to selecting drivers in a Vista setup) which also ends up doing the same thing as well.

The underlying "idea" before about needing to erase the MBR to test deploying the image was done (after my above post) using Terrabyte Lab's MBR.EXE which has a /Zero option. Anyways, after zeroing the MBR, i was able to deploy the RAID image onto the notebook, and booting resulted in the 0x7B again. But I got sidetracked... :w00t:

When I found I can set the notebook to RAID mode, which ended up that I could boot into this XP image... :o

Which then resulted in a one hour crusade to determine a way to connect a second HDD to this notebook... :sneaky:

But I haven't figured that out yet... A project for later.

Link to comment
Share on other sites

As far as the "Setup" it involved making the "ISO" (its not an ISO truly, we have the pressed disc from MS) boot from a USB key. This let us then put the iaStore.inf (and related files) into the appropriate folders as directed by our instructions. This only bypasses the "F6" method (which is just a button similar to selecting drivers in a Vista setup) which also ends up doing the same thing as well.

Well this confirms my guess that it was completely unlike "normal" (or maybe "Normal" just for you) and adds none of the info that I asked about. :(

It's not really difficult.

IF at any time during your "normal" procedure Disk Management is used, there are simply two cases (normally :whistle: ):

  1. It will see that the last two bytes of the MBR are 0000 and will prompt you to initialize the disk, etc., etc. (see above post)
  2. It will see that the last two bytes of the MBR are 55AA and will do NOTHING.

jaclaz

Link to comment
Share on other sites

After taking a look, it would appear that POSReady2009 may have a unique Setup method that is definately not exactly normal compared to XP at all!

From what I can tell, this POSReady install is just like a Vista install, including the "browse" option to find Mass Storage drivers. It's "PE" or whatever shell setup runs from does indeed support USB PnP like WinPE 2.0+ does. Looking at the disc contents, it looks to be a transition type setup that combines BOTH XP and Vista style setups. This means it does have an i386 folder and a WINBOM.ini that you would normally see on an XP install CD, but it also has a WIM file as well.

As far as checking it after every step, I don't see that as being viable, since I wasn't able to get any programs to run in it... which is why I switched to the WinPE v2.1 (x86) to run MBRFIX, since it wouldn't run in the PE that ran Setup.

I can try to do a reload tomorrow to see if I can get some behaviour from Disk Management using either Make_PE3 or even another HDD with Windows 7 on it.

Link to comment
Share on other sites

After taking a look, it would appear that POSReady2009 may have a unique Setup method that is definately not exactly normal compared to XP at all!

Maybe is the same strange approach the WINFLP has/had? :unsure:

Something half-way between Longhorn and XP rather than between XP and Vista (with a touch or two of "XP Embedded")

I vaguely remember having setup a couple of them, and they actually used if I recall correctly - in addition to F6 - a .XML :w00t: file to add specific drivers.

I am pretty sure that "basic" "no-dependencies" command line apps should be able to run in the booted Environment, just like they do with a Shift+F10 command prompt on a normal XP setup.

All in all, to diagnose all that is needed is a set of copies of the MBR, though it is possible that to take them it is needed to interrupt setup.

Something like:

dsfo \\.\PhysicalDriven 0 512 x:\somefile_m.bin

would do nicely.

jaclaz

Edited by jaclaz
Link to comment
Share on other sites

I am starting up an installation now, albeit on different hardware.. (previously mentioned notebook with RAID enabled)

The CD boots to a command prompt, where I can type "install" to begin setup. Prior to doing that, I decided to run ver which returned:

Windows 5.2.3790

After my struggles to get this installed (it isn't very fun) I will try to get a copy of the MBR for ya.

Update: I can't recreate the issue with this notebook. It is booting every time. I will try going back to the original hardware... Perhaps it works differently when there is a mirror involved. BTW all the questions it asks in setup is used to build an answer file that is used during the actual setup. Among them is an option that lists the drive type, and even if using RAID mode (at least on this notebook) it says the drive type is IDE.

Update 2: It appears the notebook itself is a root cause for some of these strange behaviours. The "RAID" setting does not operate in the same way it does on the desktop board. While it is still puzzling that the WinPE detects a broken mirror consisting of the HDD and the Card Reader, this is likely due to the fact that these Realtek Card Readers are capable of getting a drive letter without media being inserted, and that they run on the USB bus. Why specifically WinPE thinks the card reader is a 500GB drive is unknown. In addition, oddities also found are the following observations:

1. Notebook set to RAID, but XPe setup sees the HDD as IDE mode.

2. Notebook set to RAID can boot the finished RAID image without error.

3. Notebook set to AHCI or IDE will receive a 0x7B when booting the RAID image.

4. Notebook set to RAID will boot the install DVD without a problem.

5. Desktop set to RAID will receive a 0x7B when booting off the install DVD.

So it would appear that the Notebook doesn't really use its RAID function 100% while running in JBOD, or that the RAID setting isn't really a setting at all, except that it can boot XPe images that have only a RAID driver in them.

Another strange thing is that by default, on the full install of this OS, Disk Management is not included!

Either way, since this is going back to my original question regarding the inability to boot after setup, I will move this topic back into the XP forum and merge it with my original topic.

Update 3: I was finally able to reinstall POSReady 2009 and get the flashing cursor back. It is a real pain! :angry:

Here is the screenshot of Disk Management from Make_PE3, which did not prompt to format or import. It got a drive letter and all files were visible.

2011-09-01141302.jpg

See attached for first track BIN file as created by MBRWORK.

BACK0.rar

Link to comment
Share on other sites

See attached for first track BIN file as created by MBRWORK.

What you posted is "crazy".

It is partially (for the most part) a valid XP MBR BUT it has some residual of something else "below" a FAT (I presume FAT16, BUT could be FAT32) bootsector :w00t:

Second sector clearly starts with a FAT table "incipit", using F8 which means "fixed disk partiion".

In the normal XP MBR there are three "strings":

  1. Invalid partition table
  2. Error loading operating system
  3. Missing operating system

The posted MBR have them allright, but they are followed by:

  1. Invalid system disk
  2. Disk I/O Error
  3. Replace the <GARBAGE> (actually valid "DISK SIGNATURE")

These three latter strings are present in the "normal" FAT bootsector for DOS (the one invoking IO.SYS, NOT NTLDR).

This could happen if the device has been previously formatted as super-floppy as FAT and with a DOS bootsector.

One problem is the EB1A90 at offset 0xCA, which is connected with the HP USB Format tool hack:

http://reboot.pro/2246/page__st__15

The other one is the 6B2C in the reserved sectors at offset 0x1BC ( right after the disk signature).

While it is entirely possible that the EB1A90 is also something that XP Embedded (and conversely POSREADY 2009) may do, the 6B2C are really strange, I would call this a mistery^2 :ph34r: .

If you compare the MBR you posted with *any* standard MBR or with the "corrected by MBRFIX" one, you will see the above.

Was the MBR sector actually 00ed before this install attempt?

What else (which tools) have you used on the device? (in this or in in previous sessions)

From which source are you installing?

The good (or VERY bad news) on second thought, are:

The partition data of the MBR you posted is about a FAT32 partition with UNbalanced CHS/LBA values as follows:

0C-80-0-1-1-124-254-63-63-2013119

size of the partition being (LBA) 2013119x512=1,030,716,928

It means that in an access of juvenile (or senile, cannot say :unsure:) dementia :w00t: you posted the MBR of a (BADLY) partitioned/formatted 1 Gb USB stick :realmad: instead of the one of the hard disk!

:lol:

Try again.... ;)

jaclaz

Link to comment
Share on other sites

I feel silly. :blushing:

You are correct. Sadly MBRWORK was run from a 1GB USB key that boots to DOS 7.1... and it captured that MBR. Unfortunately, it does not seem to see the RAID volume.

Let me choose to "Change the Disk" and put that MBR up instead. :ph34r:

In addition I will get the MBR after MBRfix, and also the MBR of the POSReady 2009 install using IDE mode instead of RAID.

BACK1.rar

Link to comment
Share on other sites

I feel silly. :blushing:

You are correct. Sadly MBRWORK was run from a 1GB USB key that boots to DOS 7.1... and it captured that MBR. Unfortunately, it does not seem to see the RAID volume.

Well, the good news are that now you know what things are "wrong" on that USB stick ;).

Let me choose to "Change the Disk" and put that MBR up instead. :ph34r:

In addition I will get the MBR after MBRfix, and also the MBR of the POSReady 2009 install using IDE mode instead of RAID.

Good. :)

The BACK1.BIN is (as it was expected BUT NOT as it should be) a perfectly valid MBR with:

  1. NO CODE whatsoever :ph34r:
  2. Disk SIgnature allright (and the "repetitive" nature of it - A990A990 - is usually a sign hinting that it was automatically created by an "original" MS NT tool)
  3. Magic bytes 55AA allright

the question remains about HOW/WHY that happened. :unsure: really unusual.

BTW no need to post the other two MBR's, they won't help us in making more light on the issue.

jaclaz

Edited by jaclaz
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...